All the pieces together: the architecture of a web project with AI agents
An exemplary B2B portal connects infrastructure, AI agents, shared memory, security, CI/CD and operations into a coherent project.
Blog
Faster AI with controlled resources: cache, batching and quantization
AI inference optimization in cloud and on-premises: cache, batching, quantization and experiments tracking latency, memory and quality.
SEO for AI: how to make your content easy to find, understand and cite
SEO for AI, AEO and GEO: verifiable content, indexing, structured data and measurement, without guarantees of citation.
Can an AI agent use your website? A practical audit
Agentic readiness audit for the website: rendered content, semantic structure, forms, errors, and action confirmation.
Beyond “state of the art”: how do we test a programming agent
How we evaluate a coding agent beyond benchmarks: real tasks, regressions, review, and accepted results.
Where does the code come from? Dependencies, SBOM, and the security of the software supply chain
Dependencies, SBOM, and provenance: what we check before delivering code generated with AI and external software components.
Agent-ready APIs: clear actions and reliable retries
APIs for AI agents: validation, authorization, idempotency, and handling retries without uncontrolled duplication of actions.
CI/CD for agentic development: from change to controlled release
CI/CD for code written with AI: tests, verifiable artifacts, protected environments, approvals, and controlled production release.
What's happening in production? Observability for applications and AI agents
Observability for applications and AI agents: logs, metrics, traces, useful alerts, and error diagnostics without data exposure.
When the system goes down: backup, recovery, and incident response
Backup and recovery for web and AI projects: RPO, RTO, protected copies, incident response, and restoration testing.
One agent or a team of agents? How do we coordinate development without conflicts
One agent or more? Delegation, separate workspaces, handover of results, and code integration without overwrites.
The shared memory of AI agents: what we retain, who can read it, and when we forget
Agents Shared Memory explained in practice: context, provenance, access, concurrent updates, and information expiration.